Details
| Originalsprache | Englisch |
|---|---|
| Titel des Sammelwerks | ACL-IJCNLP 2021 - 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, Proceedings of the Conference |
| Seiten | 4744-4754 |
| Seitenumfang | 11 |
| Publikationsstatus | Veröffentlicht - 2021 |
| Extern publiziert | Ja |
| Veranstaltung | Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, ACL-IJCNLP 2021 - Virtual, Online Dauer: 1 Aug. 2021 → 6 Aug. 2021 |
Abstract
Generating high-quality arguments, while being challenging, may benefit a wide range of downstream applications, such as writing assistants and argument search engines. Motivated by the effectiveness of utilizing knowledge graphs for supporting general text generation tasks, this paper investigates the usage of argumentation-related knowledge graphs to control the generation of arguments. In particular, we construct and populate three knowledge graphs, employing several compositions of them to encode various knowledge into texts of debate portals and relevant paragraphs from Wikipedia. Then, the texts with the encoded knowledge are used to fine-tune a pre-trained text generation model, GPT-2. We evaluate the newly created arguments manually and automatically, based on several dimensions important in argumentative contexts, including argumentativeness and plausibility. The results demonstrate the positive impact of encoding the graphs' knowledge into debate portal texts for generating arguments with superior quality than those generated without knowledge.
ASJC Scopus Sachgebiete
- Informatik (insg.)
- Software
- Informatik (insg.)
- Theoretische Informatik und Mathematik
- Sozialwissenschaften (insg.)
- Linguistik und Sprache
- Geisteswissenschaftliche Fächer (insg.)
- Sprache und Linguistik
Zitieren
- Standard
- Harvard
- Apa
- Vancouver
- BibTex
- RIS
ACL-IJCNLP 2021 - 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, Proceedings of the Conference. 2021. S. 4744-4754.
Publikation: Beitrag in Buch/Bericht/Sammelwerk/Konferenzband › Aufsatz in Konferenzband › Forschung › Peer-Review
}
TY - GEN
T1 - Employing argumentation knowledge graphs for neural argument generation
AU - Al-Khatib, Khalid
AU - Trautner, Lukas
AU - Wachsmuth, Henning
AU - Hou, Yufang
AU - Stein, Benno
N1 - Funding Information: The first author is supported by the German Federal Ministry of Education and Research (BMBF, 01/S18026A-F) by funding the competence center for Big Data and AI (ScaDS.AI Dresden/Leipzig).
PY - 2021
Y1 - 2021
N2 - Generating high-quality arguments, while being challenging, may benefit a wide range of downstream applications, such as writing assistants and argument search engines. Motivated by the effectiveness of utilizing knowledge graphs for supporting general text generation tasks, this paper investigates the usage of argumentation-related knowledge graphs to control the generation of arguments. In particular, we construct and populate three knowledge graphs, employing several compositions of them to encode various knowledge into texts of debate portals and relevant paragraphs from Wikipedia. Then, the texts with the encoded knowledge are used to fine-tune a pre-trained text generation model, GPT-2. We evaluate the newly created arguments manually and automatically, based on several dimensions important in argumentative contexts, including argumentativeness and plausibility. The results demonstrate the positive impact of encoding the graphs' knowledge into debate portal texts for generating arguments with superior quality than those generated without knowledge.
AB - Generating high-quality arguments, while being challenging, may benefit a wide range of downstream applications, such as writing assistants and argument search engines. Motivated by the effectiveness of utilizing knowledge graphs for supporting general text generation tasks, this paper investigates the usage of argumentation-related knowledge graphs to control the generation of arguments. In particular, we construct and populate three knowledge graphs, employing several compositions of them to encode various knowledge into texts of debate portals and relevant paragraphs from Wikipedia. Then, the texts with the encoded knowledge are used to fine-tune a pre-trained text generation model, GPT-2. We evaluate the newly created arguments manually and automatically, based on several dimensions important in argumentative contexts, including argumentativeness and plausibility. The results demonstrate the positive impact of encoding the graphs' knowledge into debate portal texts for generating arguments with superior quality than those generated without knowledge.
UR - http://www.scopus.com/inward/record.url?scp=85118945370&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85118945370
SN - 9781954085527
SP - 4744
EP - 4754
BT - ACL-IJCNLP 2021 - 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, Proceedings of the Conference
T2 - Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, ACL-IJCNLP 2021
Y2 - 1 August 2021 through 6 August 2021
ER -